create_training_data: Create training data for GPT

In arete: Automated REtrieval from TExt

Create training data for GPT

Description

Open WebAnnoTSV files following RECODE structure and build training data for large language models in a variety of formats.

Usage

create_training_data(
  input,
  prompt = NULL,
  service = "GPT",
  aggregate = TRUE,
  export_type = "jsonl",
  out_path = NULL
)

Arguments

input	character or list. Either a set of paths to WebAnno TSV 3.3 files from which the text and annotated data are taken or a list with two terms, 1) paths to .txt or .pdf files e.g: "./folder/file.pdf" from which text data will be taken from and 2) paths to WebAnno TSV 3.3 files from which to take annotation data.
prompt	character. Custom prompt to be attached to each text during construction of the training data. Default prompt used otherwise.
service	character. Service to be used. Right now, only GPT is available.
aggregate	boolean. If TRUE and prompt is "csv", a single csv is created.
export_type	character. Either "jsonl" or "csv". If "jsonl", a single file is created in which each line is a json specifying the input (prompt and text) and expected output (data).
out_path	character. Path to where the training data will be saved.

Value

matrix / data.frame

Examples

example = system.file(paste0("extdata/insecta_annot_1.tsv"), package = "arete")

create_training_data(input = example, service = "GPT", export_type = "jsonl")

arete documentation built on Nov. 5, 2025, 6:31 p.m.